Compared with the Conventional MR Imaging, Do the Constructive Interference Steady State Sequence and Diffusion Weighted Imaging Aid in the Diagnosis of Lumbar Disc Hernias?

Abstract

Objective:

Our aim was to evaluate the efficacy of Diffusion Weighted Imaging (DWI) and Constructive Interference Steady State (CISS) sequences in the depiction of lumbar disc hernias, including sequestrated and extruded discs.

Materials and Methods:

100 patients; 35 males and 65 females, with 3 sequestration-16 extruded and 81 protruded all together 100 herniated discs were involved in this study. If a patient had more than one herniated disc, then the more precise herniation was included. All the MRI procedures were handled in Siemens Symphony Power 1.5 T. magnet. DWI- Apparent diffusion coefficient (ADC mapping) and CISS sequences were obtained in axial planes. Statistical analysis of CISS and DWI results were analysed by fischer’s test, chi square statistics.

Results:

CISS sequence had a 89% sensitivity, 100% specificity for all lumbar disc hernias, had 100% sensitivity and specificity in the evaluation of sequestrated and extruded discs. DWI had a 77% sensitivity and %100 specificity, had 100% sensitivity and 50% specificity in the evaluation of extrusion, 100% sensitivity and specificity in the evaluation of sequestration.

Conclusion:

CISS sequence and quantitative DWI may be alternative imaging modalities to the routine MR imaging sequences in the depiction of lumbar disc hernias, including sequestration and extruded discs.

Introduction

The most common cause of back pain is the herniation of an intervertebral disc, which may cause nerve impingement and radicular symptoms or a neurological deficit [1, 2]. Degenerative diseases of the lumbar spine and disc herniations most frequently call for evaluation and diagnosis through imaging [1, 3]. Magnetic Resonance (MR) imaging is considered to be the method of choice for the depiction of intervertebral disc pathologies [1, 3, 4]. For the analysis of lumbar disc hernias, conventional lumbosacral MR imaging and MR myelography can provide adequate information, but recent developments in MR hardware technology and pulse sequences have further refined the technique [1, 4]. Constructive Interference Steady State (CISS), a gradient-echo technique with steady-state free precession, and Diffusion Weighted Imaging (DWI)-Apparent Diffusion Coefficient (ADC) mapping are newer pulse sequences that have been proposed to improve the quality of routine MR imaging modalities [1, 4, 5].

In this study, we illustrate and discuss the utility of 3-Dimensional (3D)-CISS and DWI-ADC mapping in the diagnosis of lumbar intervertebral herniations, including extruded and sequestrated discs, as compared with the MR imaging modality.

Materials and Methods

This study consisted of 100 consecutive patients referred to our institution with back pain and radicular symptoms. From May to November 2009, 35 males and 65 females were analyzed who ranged in age between 17 and 77 years with a mean of 47 years. Among the patients in this study, there were 3 sequestrated discs, 16 extruded discs and 81 protruded discs. For patients with more than one herniated disc at different lumbar intervertebral levels, more precise visualization of the herniation was included. Informed consent was obtained from all patients prior to imaging, and the study was approved by the ethics committee. Surgical and histopathological findings were considered the gold standard for diagnosis of the disc herniations that were treated through surgery, and routine lumbar MR imaging was performed for the nonoperated herniations.

MR examinations were performed with a Siemens Symphony Power 1.5 T. magnet (Siemens-Erlangen-Germany) with 30-mlt/m maximum gradient strength and a 120-mt/mper-millisecond slew rate using a standard body and spine coil. The conventional lumbar MR imaging protocol included T1 and T2W TSE sagittal and T2W TSE axial planes. The parameters of the 3D-CISS sequence were as follows: TR/TE, 5.64/2.82; FOV, 225*300; slice thickness, 1 mm; NEX, 2. The flip angle was 55 degrees, and the scan duration was 2.35 min. DWIs were obtained by 3D-SE Echoplanar Imaging (EPI) with TR, 960; TE, 89; FOV, 300*335 mm; acquisition time, about 0.25 min. ADC mapping was also conducted with b values of 50 and 400 s/mm². Both the 3D-CISS and 3D-DWI sequences were handled and analyzed in the transverse planes.

Analysis of the data set

All images were interpreted by a single neuroradiologist with 8 years of experience. Herniations with poor image quality were excluded from the study. The following terms were used to describe disc abnormalities: normal, bulging, protruding, extruding and sequestered. A disk was considered to be normal when it did not reach beyond the borders of the adjacent vertebral bodies, and bulging was defined as a circumferential symmetric disc extension around the vertebral border [6, 7]. Protrusion was defined as a focal or broad-based extension of the annulus with smooth margins beyond the vertebral margin with the base of the disc origin broader than the apex and an intact outer annulus/ posterior longitudinal ligament complex [6, 7]. Extrusion was defined as a more extreme extension of the disc beyond the vertebral border and a narrow base against the disc of origin as well as a disrupted outer annulus/posterior longitudinal ligament complex. Sequestration was defined as a disc fragment that had been broken off from the parent nucleus and was free in the epidural space [6, 7]. The protruded, extruded and sequestrated discs were designated as herniated discs; normal and bulging discs were excluded from the study. If present, nerve roots and thecal sac compressions were also examined. There were also annular tears, osteophytes, disc and facet joint degenerations, spinal stenosis and so forth in some of our patients.

The statistical analysis of the CISS and DWI results from the MRI and histopathological findings were performed by Fischer’s test and Pearson chi square statistics using the SPSS 15.0 written form (SPSS-Inc, Chicago-IL). p>0.05 was considered to indicate a statistically nonsignificant difference.

Results

A total of 3 sequestrated, 16 extruded and 81 protruded discs were included in this study; 20 of the diseased intervertebral discs were submitted to open surgery and were histopathologically confirmed. All of the sequestrations, 10 of the 16 extrusions and 7 of the 81 protrusions were surgically repaired.

The CISS sequence correctly diagnosed 88 disc injuries, and DWI-ADC mapping with b values of 50 and 400 s/mm² identified 76 herniated discs. All of the extruded and sequestrated discs were evaluated using the CISS sequence. Only one extruded disc was undetected by DWI and ADC mapping.

When we compare the CISS sequence results with those obtained through routine MRI, CISS demonstrated 89% sensitivity and 100% specificity for all lumbar disc hernias (p=0.120) (Table 1, 2 and Figure 1, 2), as well as 100% sensitivity and specificity for the detection of sequestrated and extruded discs (p=0.059) (Table 3, 4 and Figure 3, 4). The CISS sequence results were not statistically different from those of routine MRI for the herniated, sequestrated and extruded discs (Fischer’s test, p>0.05). The positive predictive value (PPV) and negative predictive value (NPV) of the CISS sequence for herniations were 1.00/0.08, and for sequestrations and extrusions, the values were 1.00/0.00.

Table 1.

CISS sequence findings in comparison with routine MR imaging

			MRI Findings		Total
			Negative	Positive
CISS Sequence Findings	Negative	n	1	11	12
		n%	8%	82%	100%
	Positive	n	0	88	88
		n%	0%	100%	100%
Total		n	1	99	100
		n%	1%	99%	100%

p=0.120, Sensitivity=0.89, Specifity=1.00, PPV=1.00, NPV=0.08

Table 2.

DWI-ADC map findings in comparison with routine MR imaging

			MRI Findings		Total
			Negative	Positive
DWI-ADC Map Findings	Negative	n	1	23	24
		n%	4%	96%	100%
	Positive	n	0	76	76
		n%	0%	100%	100%
Total		n	1	99	100
		n%	1%	99%	100%

p=0.240, Sensitivity=0.77, Specifity=1.00, PPV=1.00, NPV=0.04

Figure 1.

CISS sequence findings in comparison to routine MR imaging.

Figure 2.

Lumbar disc herniation shown in T2W sagittal (A), T1W sagittal (B), T2W axial (C) images and CISS sequence (D) in axial plane.

Table 3.

CISS sequence findings for extruded discs in comparison with routine MR Imaging. In our study sensitivity and spesivity of CISS sequence for extrusion was 100%

			CISS Sequence Findings		Total
			Negative	Positive
Extrusion	Negative	n	1	0	24
		n%	100%	0%	100%
	Positive	n	0	16	16
		n%	0%	100%	100%
Total		n	1	16	17
		n%	6%	94%	100%

p=0.059, Sensitivity=100%, Specifity=100%, PPV=100%, NPV=0

Table 4.

DWI-ADC map findings for extruded discs in comparison with routine MR Imaging. In our study sensitivity and spesivity of CISS sequence for extrusion was 100%

			DWI-ADC Map Findings		Total
			Negative	Positive
Extrusion	Negative	n	1	0	1
		n%	100%	0%	100%
	Positive	n	1	15	16
		n%	6%	94%	100%
Total		n	2	15	17
		n%	12%	88%	100%

p=0.118, Sensitivity=100%, Specifity=50%, PPV=94%, NPV=0

Figure 3.

CISS sequence findings for extruded discs in comparison with routine MR Imaging. In our study, Sensitivity and spesificity of CISS sequence for extrusion, were 100%.

Figure 4.

T2W sagittal image (A), shows sequestration of the disc. Axial T2W image (B) and CISS sequence in axial plane (C) show the indentation of the dural sac by the sequestrated fragment.

In contrast to routine MRI findings, DWI and ADC mapping demonstrated 77% sensitivity and 100% specificity for all hernias (p=0.240) (Table 1, 2; Figure 5, 6), as well as 100% sensitivity and 50% specificity for the extrusions (p=0.118) (Table 3, 4; Figure 7, 8) and 100% sensitivity and specificity for the sequestrations. Similarly to the CISS sequence, DWI and ADC mapping showed no statistically significant differences compared with the conventional MRI findings for any of the injured discs (Fischer’s test, p>0.05). The PPV and NPV of DWI and ADC mapping for lumbar hernias were 1.00/0.04. For extrusions, the values were 0.94/0.00, and for the sequestrations, the values were 1.00/0.00.

Figure 5.

DWI-ADC map findings in comparison with routine MR imaging.

Figure 6.

T1W sagittal image (A), shows extrusion at L5-S1 level and dural sac indentation is clearly seen in axial plane, CISS sequence (B). ADC map (C) and DWI (D) of the same patient are also depict extrusion in axial planes (arrows).

Figure 7.

DWI-ADC map findings for extruded discs in comparison with routine MR Imaging.

Figure 8.

CISS sequence in axial plane (A) indicates herniation in the central and left paracentral zone. ADC map (B) and DWI (C) in axial plane also indicates the herniation at the same location as seen with arrows.

When comparing the CISS and DWI results, we observed no statistically significant difference between the sequences obtained among all of the disc herniations (p>0.05). A sensitivity of 91% and a lower specificity of 21% were observed at a p value of 0.153. The PPV and NPV for both sequences were 0.78/0.42 (Table 5, Figure 9).

Table 5.

Concordance of CISS and DWI findings: No significant statistical differences of sequences from each other for the evaluation of all disc herniations (p>0.05)

			DWI-ADC Map Findings		Total
			Negative	Positive
CISS Sequence Findings	Negative	n	5	7	12
		n%	42%	58%	100%
	Positive	n	19	69	88
		n%	22%	78%	100%
Total		n	24	76	100
		n%	24%	76%	100%

p=0.153, Sensitivity=0.91, Specifity=0.21, PPV=0.78, NPV=0.42

Figure 9.

Concordance of CISS and DWI findings: No significant statistical differences of sequences from each other for the evaluation of all disc herniations (p>0.05).

When we combined the results of both sequences for all lumbar disc hernias, we did not detect any statistically significant differences from the routine MR findings (p>0.05). A sensitivity of 70% and specificity of 100% were observed at a p value of 0.310. The PPV and NPV of the combined sequences were 1.00/0.03 (Table 6, Figure 10).

Table 6.

Combination of results of both sequences for all lomber disc hernias: No significant statistical differences of the combination from the routine MR findings either (p>0.05)

			MRI Findings		Total
			Negative	Positive
CISS –DWI Combination	Negative	n	1	31	12
		n%	3%	97%	100%
	Positive	n	0	69	88
		n%	0%	100%	100%
Total		n	1	99	100
		n%	1%	99%	100%

p=0.310, Sensitivity=0.70, Specifity=1.00, PPV=1.00, NPV=0.03

Figure 10.

Combination of results of both sequences for all lomber disc hernias: No significant statistical differences of the combination from the routine MR findings either (p>0.05).

Discussion

Back pain is one of the leading causes of disability and absence from work. Disc abnormalities in most patients may be the primary cause of these symptoms [2, 7]. A number of imaging modalities are available for the assessment of disc injuries, such as X-ray and Computerized Tomography (CT)-myelography. Over the last 10 years, MR imaging of the spine has shown significant advances and have evolved to be the method of choice for the depiction of lumbar disc hernias [2, 6–8]. The sensitivity and specificity of MR imaging for inter-vertebral disc abnormalities are relatively well known [7, 8]. Because of improvements in the coil technology, more rapid scanning and high-resolution imaging sequences have led to significant advances in the imaging modalities used for the diagnosis of spinal pathologies [6, 7]. Here, we describe the use of two new MRI sequences, 3D-CISS and DWI-ADC mapping, with respect to their reliability-utility and efficacy for the analysis of lumbar disc hernias.

DWI and ADC mapping allow us to observe the mobility of water molecules within the tissues. The signal intensity of a mobile water molecule decreases as the b factor increases, while that of an immobile water molecule does not [5, 9]. Therefore, the CSF-containing thecal sac and the nerve roots will show low signal intensities on DWI, while solid and semi-solid structures, such as herniated disc fragments, will show increased signal intensity allowing for differentiation of the fluid from solid components [5, 9]. The disadvantage of less spatial resolution is compensated by the correlation of CISS images [4, 5]. DWI and ADC mapping are routinely used in brain studies, and these techniques are important for identifying ischemic and neoplastic lesions. Currently, these sequences are also used in spinal, abdominal and musculoskeletal pathologies [9].

In this research, we have applied b values of 50 and 400 s/mm² for the ADC mapping. As these values are lower than 500 s/mm² and close to the T2W images with a b value of “0”, we decided to obtain high-quality images with higher spatial resolution. We chose the better image with b values of 50, 400 s/mm², or we used both in some cases.

CISS sequence: The 3D-steady-state refocused gradient-echo sequence is flow-compensated, high-contrast and high-spatial resolution with the slice thickness reduced to 0.7 mm, resulting in detailed visualization and differentiation of the various components of the nerve roots, vessels, CSF and the neural structures surrounded by CSF [4, 5, 10]. Therefore, this sequence has been mostly used in the assessment of cranial nerves to evaluate their pathologies in the cerebellopontine angle and inner ear [5, 10–12]. As a short TE and low flip angle is employed, the CISS sequence is amenable to 3D acquisition, high T2/T1 and high CSF/nerve contrast and accentuated T2 weighting. Therefore, the CISS sequence can provide a myelographic image with excellent CSF-to-spinal cord contrast and visualization of the thecal sac and dural root sleeves at each intervertebral level [4, 10, 13]. Its disadvantage, similar to other gradient echo sequences, is its high sensitivity to motion. Therefore, flow can easily result in a loss of signals due to pulsatile CSF motion. However, CISS variants, such as Free Induction with Steady-State precession (FISP) pulses and compensation, can be applied over each TR cycle, making this modality quite suitable for imaging the neural structures [4, 10]. The sequence cannot properly differentiate discs from osteophytes, but it can easily reveal the site of injury and demonstrate the effects on the thecal sac and spinal cord [1, 4]. The thinner slices and reduced partial volume effect of the CISS sequence also enable suitable visualization of spinal and neuroforaminal canals in patients with spondylosis [4, 13].

In the literature on this topic, we found only a single article devoted to CISS sequence. Maksymowich et al. [1] studied CBASS (Completely Balanced Steady State sequence), which is synonymous with the CISS sequence in low-field open MR systems. The authors analyzed degenerative diseases of the lumbar spine in 78 cases. They showed that the CBASS sequences were superior to T2W sequences for depicting the disc herniation and the relationship of the herniated fragment to the dural sac, nerve roots and bulged discs. However, CBASS was not superior for the evaluation of vertebral, facet joint or disc degeneration, spinal canal stenosis or herniated discs with respect to the relationship of the posterior longitudinal ligaments. Their report did not evaluate DWI or ADC mapping. Our CISS sequence, conducted in a highfield MR system, was also highly sensitive and specific in the evaluation of disc herniations, especially for the extruded and sequestrated discs. We also had some annular tears, osteophytes, disc and facet joint degenerations, spinal stenosis and so forth among our patients, but we did not focus on these injuries. The CISS sequence with higher-quality imaging provided the data necessary for the depiction of lumbar disc hernias, and this technique may be used as an alternative imaging modality rather than the DWI. Although these sequences had no statistical differences compared with routine lumbar MR imaging and had slightly lower sensitivity and specificity than the conventional MRI, we present these two imaging modalities, particularly the CISS sequence, to provide new imaging techniques. We sought to analyze the applicability of the CISS sequence for the depiction of disc herniations.

To our knowledge, this is the first paper to use CISS and DWI-ADC mapping for the analysis of lumbar disc herniations. Both sequences had higher specificity for defining the real disc herniations. The CISS sequence had substantially higher sensitivity than DWI-ADC mapping (89%/77%) for the imaging of all lumbar disc hernias. For the diagnosis of the extruded and sequestrated discs, the CISS sequence had the highest sensitivity and specificity (100%) whereas DWI-ADC mapping had similar results for sequestrations but lower specificity for the extruded discs (50%). The sequences had high PPV and low NPV values, and the results of both sequences were statistically concordant with each other, demonstrating a low specificity and higher sensitivity. The combination of the CISS and DWI sequences was also useful in the diagnosis of lumbar disc herniations, demonstrating 70% sensitivity and 100% specificity.

For the 20 patients who received surgical operations, surgery was considered to be the gold standard for the diagnosis of spinal injury; for the remaining 80 patients, routine lumbar MR imaging was taken into account. In this study, only 20% of the patients had surgical and histopathological verification, and these were mostly represented by those with sequestrated and extruded discs. The other 80 patients did not undergo surgical processes due to their refusal. This was the most important limitation of our research. As there was only one special neuroradiologist with experience in this field in our department, there was no intra- or interobserver match. This may be a second important limitation of this study. The total imaging time for both sequences, including ADC mapping with the two b factors described above, was about 3 min. Therefore, the application of either sequence could provide new diagnostic information without any significant prolongation of the scan time.

We propose that transverse images of 3D-CISS and DWI sequences could be included in the routine diagnostic lumbosacral MRI protocol.

In conclusion the CISS sequence and quantitative DWI may be alternative imaging modalities to routine MR imaging sequences for the depiction of lumbar disc hernias, including sequestration and extruded discs. The 3D-CISS sequence has advantages compared with DWI-ADC mapping for the assessment of disc herniations, but either method alone or the combined use of both sequences will provide important data for the diagnosis and management of lumbar intervertebral disc injuries.